Kawalan ibu bapa yang baru dan rsquo; s dan perlindungan kesihatan mental untuk ChatGPT bertujuan untuk menjadikan AI lebih selamat untuk remaja.

• OpenAI rolls out parental controls for ChatGPT
• Parents can now connect with their teen’s accounts, limit features, and get alerts if emotional distress is detected
• High-risk conversations will be handled by specially tuned models trained to support users in crisis

OpenAI is launching a suite of safety enhancements for ChatGPT aimed at better protecting teenage users and individuals experiencing emotional hardship. The company revealed new parental supervision tools that allow parents to link their accounts with those of their teens (aged 13 and up). These tools will enable caregivers to set usage boundaries and receive instant notifications if the AI identifies potential signs of depression, anxiety, or emotional crisis in a conversation.

This move signals OpenAI’s acknowledgment that teens are actively using ChatGPT—often as a confidant or emotional outlet. While not explicitly stated, the update appears to be partly influenced by recent, widely publicized cases where families blamed AI chatbots for contributing to mental health crises, including suicides.

The rollout will begin within the next few weeks. Once a parent links their account, they can control whether chat history is saved or if the AI’s memory function is active. Default settings will include age-appropriate content filters. Parents will only be alerted when the system detects serious concerns—ensuring privacy is maintained in everyday use, while allowing intervention during critical moments.

“In our ongoing effort to make ChatGPT as supportive and safe as possible, we’ve learned that people often turn to it during deeply personal struggles,” OpenAI shared in a recent blog post. “That’s why we’re continuously refining how our models detect and respond to emotional distress—with guidance from leading experts.”

Specialized Models for Sensitive Conversations

For both teens and adults, OpenAI will now redirect conversations involving mental health challenges or suicidal thoughts to a dedicated version of its AI model. This variant uses a technique known as deliberative alignment to respond more carefully, avoid harmful influences, and remain firmly within safety protocols—even under manipulative or high-pressure prompts.

To support this system, OpenAI has formed the Expert Council on Well-Being and AI and the Global Physician Network, a coalition of over 250 healthcare professionals specializing in mental health, addiction, and adolescent development. These experts will help define how the AI identifies distress, shapes compassionate responses, and determines when real-world intervention may be necessary.

While parental concerns have traditionally focused on screen time and online content, AI introduces a new dimension: the nature of the interactions happening behind the screen. When a teenager confides in an AI that mimics empathy and understanding—despite being code—it blurs emotional lines and raises new risks.

Until now, AI safety measures have largely been reactive. These updates mark a shift toward proactive harm prevention. The hope is that instead of waiting for a crisis to escalate into a dire message or a desperate plea from the AI to a parent, early detection can gently redirect someone in distress toward support—before it’s too late. It may feel intrusive to some, but if it helps pull someone back from the edge, the discomfort may be worth it.

Atas ialah kandungan terperinci Kawalan ibu bapa yang baru dan rsquo; s dan perlindungan kesihatan mental untuk ChatGPT bertujuan untuk menjadikan AI lebih selamat untuk remaja.. Untuk maklumat lanjut, sila ikut artikel berkaitan lain di laman web China PHP!